[Clascomment] OPT-IN: phi-meson photoproduction on Hydrogen in the neutral decay mode

[Marco Ripani]

Dear all,
indeed in the first round of comments from our Ad Hoc committee there 
was a question about the sentence on the failure of the fit of the mass 
distribution at very backward angles. We were asking whether you meant 
that the high chi^2 was indicative of statistical inconsistency. In 
other words, we were guessing it simply meant that for some reason your 
parameterizazion of the phi mass distribution and the background fails 
to reproduce the data at those angles (by the way, following Volker's 
comment on the Voigtian function, may this be due to resolution effects 
becoming more important at very backward angles ?).
The committee asked many other questions and made many other comments in 
multiple rounds, which required important and extensive changes in the 
paper. Moreover there was some pressure in trying to align the 
collaboration review of this paper with that of the CMU paper, so that 
somehow that particular point got out of sight and was left in the paper.
I think it should simply be changed in "at very backward angles our 
parameterization fails to reproduce the data, as indicated by the high 
value of chi^2/dof" (indeed explicitly reporting the chi^2 values in the 
graphs would help assessing the goodness of the fit).

Regarding the acceptance cut and its effect on the data, the committee 
also discussed this point but I prefer to leave it to the lead authors 
to clarify this point.

Best Regards
Marco

Il 16/05/2013 19:08, Volker Burkert ha scritto:
> Dear Heghine and Moscov,
>
> This is an important paper. We like to make it as good as possible. Here are a few comments that
> you should consider.
>
> Introduction:
>   Left column: line 6: The statement is too strong. There are model analyses by J.J. Xie, B.S. Zou, H.C. Chiang,
> Phys.Rev.C77 (2008) 015206 that discuss evidence for N*(1535) ->N-phi coupling, and a bit more general
> in terms of qqqs-sbar contributions in:  C. S. An, B. S. Zou, Eur.Phys.J. A39 (2009) 195. These papers should
> be referenced.
> You may also say a few words with references to the CLAS phi-electroproduction publications, and in case
> the charged channel phi paper makes it to the arXiv's in time, reference it as well.
>   
> Background selection.
> As a signal function for the background evaluation a Breit-Wigner form is used with the width of the
> phi (=4.3MeV). However, the actual signal “width” appears to be more like 15-17MeV in Fig. 5, which means
> that the width is dominated by resolution effects. Therefore the fit should include a function convoluted
> of a Gaussian and a Breit-Wigner form, also known as a Voigt profile (http://en.wikipedia.org/wiki/Voigt_profile).
> Further down it says: “The fit at the very backward angle has very large chi^2 due to the limited statistics
> at large angles”.  I assume you mean the reduced chi^2. The statement makes no sense as the reduced
> chi^2 = chi^2/#DF, should be independent of statistics. In any case you should  give the chi^2/#DF values.
>
> Systematic uncertainties.
> The acceptance cut is either not well described, or not justified. If I understand it correctly, it may bias
> your data. Please describe the procedure in more detail and explain why it does not bias the results.
>
> Figures.
> Fig. 7 and Fig. 8. The high density of the data combined with the large symbol size makes the graphs difficult
> to read. You may want to reduce the symbol size.  Also, the systematic error histograms dominate the appearance
> of the graphs. You may use a light grayish pattern or just unfilled white histograms.
> Fig. 9, 10.  I have the same comment on the systematic error histograms. Use a light grayish pattern or
> just unfilled white histograms.
> Fig. 11, 12.
> The data points at large energies and backward angles are impossible to distinguish from the zero line.
> In order to make the data useful for the reader I suggest to tabulate the cross sections for all bins. There
> are about 20 energy bins and 14 angle bins, i.e. 280 data points that could easily be accommodated in a
> few extra pages.  There one should also include the statistical and systematic errors.
> Fig. 13.
> I suggest you use different symbols for the CLAS data, e.g. red-filled black circles that are slightly larger
> than the other symbols, the latter should be shown as open symbols. The CLAS data should dominate
> the appearance as they represent a more complete set covering the entire bump structure as well as
> the transition to high energy.
>
> Volker
>

-- 

Marco Ripani
Senior Scientist, Istituto Nazionale di Fisica Nucleare - Genova - Italy
Address:
INFN
Via Dodecaneso 33
16146 GENOVA - ITALY
tel./fax +39-010-3536458
cell. +39-328-2191138
E-Mail: ripani at ge.infn.it